AI Agents with Python and CrewAI

Large Language Models can answer questions, but they can't act on their own. They can't search the web for current information, query a database, send an email, or trigger a notification. This course bridges that gap. You will learn how to build autonomous AI agent systems that don't just think — they do.

We start with the foundations: what milestones shaped modern AI, how we got from simple language models to agentic AI, and what distinguishes an AI agent from a chatbot. You will understand the key concepts — roles, goals, tools, tasks, and crews — and learn why multi-agent orchestration is one of the most powerful design patterns in applied AI today.

The core of this course is CrewAI, a Python framework for building collaborative multi-agent systems. You will go from zero to fully functional agent pipelines through three hands-on use cases, each building on the previous one:

In Use Case 1 (Tool Integration), you experience the knowledge cutoff problem firsthand, then solve it by giving agents access to web search via SerperDevTool. You also build your first custom tool — an email sender using the Brevo API — and orchestrate a two-agent system where one agent researches and another sends the results by email.

In Use Case 2 (Database Monitoring & Automated Reporting), you build custom tools for SQLite queries and Excel report generation. Two specialized agents work together in a sequential pipeline: a Database Specialist identifies products with low inventory, and a Reporting Specialist generates an Excel report and sends an email notification. This use case also demonstrates working with alternative LLMs like DeepSeek.

In Use Case 3 (Web Scraping & Multi-Channel Notifications), you build a three-agent system that scrapes book data from a website using Selenium, sends a professional summary email, and triggers an SMS confirmation. This is a complete multi-agent, multi-channel automation pipeline.

By the end of this course, you will know how to define agents with specific roles, build custom tools that connect to real APIs and databases, orchestrate multi-agent workflows, and integrate external services like web search, email, databases, web scraping, and SMS into your agent systems. Everything is taught through practical, runnable Jupyter Notebooks that you can adapt to your own projects immediately.

No prior experience with AI agents is required — basic Python knowledge is all you need to get started.